conclusion.tex

\chapter{Future Work}
\label{sec:futureWork}
There are many different directions that researchers could follow to build off of this work.  First, when I was fuzzing, I only consistently tested after each input if the application completely crashed.  However, the one bug that I did find, crashing the \texttt{VIM} extension in \texttt{VS Code}, was only found by using the application while I was fuzzing it.  I did attempt to use each application while I fuzzed it---playing music from \texttt{Spotify}, writing code in \texttt{VS Code}, and using the Internet while fuzzing \texttt{mDNSResponder}---but I did not do this in an organized and reproducible fashion, nor did I do it the entire duration of fuzzing.

I would also like to look at how these applications work by tracing them as well.  I was unable to acquire a computer that I had \texttt{sudo} access to and that I felt comfortable turning off System Integrity Protections, so I was unable to trace any application.  By tracing an application, I would be able to see the exact system calls and arguments used for each system call, and get a better understanding of exactly what each application was doing.  This could let me have a more targeted approach to my fuzzing.

Finally, I want to look deeper into the BrokenPipeErrors that I received when fuzzing \texttt{VS Code} and \texttt{launchd}.  I hypothesize that these errors are evidence of security features that help prevent against fuzzing and attempts to find input-based vulnerabilities.